2512
VAN L. BOHNSON AND A. C. ROBERTSON
VOl. 45
3. The partial molal free energy of mercury in some mercurides has been calculated from the measured values of the electromotive force of amalgam concentration cells. 4. The values for the partial molal free energy of mercury in sodium and lead mercurides were found to be -61 and -12 calories, respectively. 5. The free energies of formation of sodium and lead mercurides are -18,351 and -280 calories, respectively. 6. The partial molal free energies of the constituents of mixed crystals of zinc-mercury, cadmium-mercury and thallium-mercury crystals have been tabulated. 7. With the exception of thallium, it has been found in the case of the mercurides and the mixed crystals studied, that the change in free energy has been greater for the electropositive metals than for mercury. CAMBRIDGE 39, MASSACHUSETTS [CONTRIBUTION FROM THE LABORATORIES OF GENERAI,CHEMISTRY OF THE UNIVERSITY OF WISCONSIN]
PROMOTER ACTION IN HOMOGENEOUS CATALYSIS. I. COPPER SALTS AS PROMOTERS IN THE IRON SALT CATALYSIS OF HYDROGEN PEROXIDE BY VANI,. BOHNSON AND A.
c. ROBERTSON
R E C E I V JUNE ~ D 11, 1923
The investigation described in this paper was an outgrowth of an earlier study1i2of the catalytic decomposition of hydrogen peroxide, during which i t was noted that the rate of evolution of oxygen in the presence of a mixture of iron and copper salts is much more rapid than the rate which may be calculated on the assumption that each salt acts independently. This phenomenon is analogous to what has been known in cases of heterogeneous catalysis as “promoter action.” The latter designation, first used in the patents of the Badische Anilin und Soda Fabrik, might well include, according to Pease and T a y l ~ rall , ~those cases in which a mixture of two or more substances is capable of producing a greater catalytic effect than can be accounted for on the assumption that each substance in the mixture acts independently. So defined, the term is applicable not only to cases of heterogeneous catalysis, to which it was first applied, but also to homogeneous catalysis, and in this sense i t is used in this paper; such a use is no intimation that the mechanism of reaction in the two different systems is the same. The known examples of promoter action have been reviewed and classified by Pease and T a y l ~ r and , ~ therefore will not be described here. Ex1
Bohnson, J. Phys. Chem., 25, 19 (1921).
8
Pease and Taylor, J. Phys. Chem., 24,241 (1920).
* Bohnson and Robertson, THISJOURNAL, 45,2493 (1923),
NOV.,
1923
PROMOTER ACTION IN HOMOGENEOUS CATALYSIS,
I
2513
amples in the field of homogeneous catalysis are relatively few in number, the one best known being the effect of neutral salts in accelerating the acid-catalyzed hydrolysis of esters. Two reactions which are of interest in connection with the present study are the oxidation of potassium iodide by potassium persulfate,* and the reaction between hydrogen peroxide and hydriodic acid;s in both of these cases the reaction was “promoted” by the action of copper salts in the presence of iron salts.
Purpose and Method In an attempt to outline in a preliminary way the mechanism of promotion, this paper will be devoted to a discussion of the factors involved in the accelerating effect of copper salts upon the iron-salt catalysis of hydrogen peroxide in acid solution. Further papers will deal with promoter action in other homogeneous reactions, such as the activation of chromic acid and its salts (catalyzing the decomposition of hydrogen peroxide) by certain cations, a reaction which is now being studied in this Laboratory. It has been suggested by Pease and Taylor3 that a distinction be made between simple “activation” of one catalyst by another substance not itself a catalyst for the reaction in question, and (‘co-activation,” in which the catalytic effect of each active substance is increased by their joint presence. From this point of view, it is possible with the reaction in question that (1) the copper salt may be activated by the iron salt, (2) the iron may be activated by the copper, or (3) the two may activate each other. It is not always easy to decide among such hypotheses, but a determination of .the promotion effect in the presence of various amounts of iron or copper may enable us to exclude certain of them, on the reasonable expectation that in any activation of one substance by another there is a limit to the accelerating effect of the primary catalyst when all the catalytic material has been “activated.” The method used for measuring the rate of reaction has been previously described.2 The results obtained are merely summarized as in the previous paper, and are supported by numerous duplicates and intermediate values. In the tables the following abbreviations are used: T , temperature; N , normality of free acid present; C,,, concentration of iron salt in milligram-atoms of iron per liter; Ccu, concentration of copper salt in milligram-atoms of copper per liter; KR,the relative rate of reaction (see p. 2497 of preceding paper). In all experiments, unless otherwise noted, the concentration of free acid was 0.076 N , that of the hydrogen peroxide 130 millimoles per liter, and the temperature 30”. All experiments were performed in the diffused light of the laboratory, after it was 5
Price, 2.phys. Chem., 27,499 (1893). Brode, ibid., 37, 257 (1908).
2514
VAN L. BOHNSON AND A. C. ROBERTSON
VOl. 45
found that those performed in darkness were not retarded either in the presence of iron alone or of iron and copper together. Preliminary Experiments Ferric chloride in solution exerts an energetic decomposing effect on hydrogen peroxide even in the presence of hydrochloric acid, which retards the catalysis materially. Ferric sulfate in sulfuric acid solution has a marked effect, less than that of the chloride. On the other hand, much larger quantities of cupric chloride, either in acid or in neutral solution, have very little noticeable effect. In fact, the salts of few other metals act as vigorously as the iron salts. When, however, small amounts of copper salt are mixed with the iron salt, the rate of reaction is very markedly increased. This effect seems to be peculiar to copper compounds; nickel, cadmium, mercury, barium, lead, cobalt, zinc, tin, and magnesium salts appear to retard rather than accelerate the reaction. I n sulfate systems silver appears to have a very slight accelerating effect. Effect of a Fixed Concentration of Copper in the Presence of Variable Iron Concentration Fig. 1shows graphically the relation between the rates of reaction caused by each salt alone and by mixtures containing various fixed concentrations
Fig. l.-Effect
Mg.atoms per liter of Fe as Fee13 on reaction rate of fixed copper with varying iron concentration
of copper salt and varying concentrations of iron salt; the data for the curves are given in Tables I and 11. In all these experiments the chlorides were used; the effect in sulfate systems will be cQns.idere$ later.
Nov., 1923 PROMOTER ACTION IN HOMOGENEOUS CATALYSIS. I
2515
TABLE I COMPARISON OF CATALYTIC EFFECT OF FERRIC AND COPPERCHLORIDES T = 30' N = 0.076 CFe ccu KR CFs CCu KR cF0 CCu KR 20 0 177 7.5 0 69 32.5 0.63 0 0 195 0 100 22.5 1.16 10 0 37.5 32.5 0 248 0 118 1.8 12.5 50 0 37.5 0 276 17.5 0 160 0 19.5 2.5
TABLE I1
EFFGCT ON THE RATEOF REACTION OF VARYING THB FERRIC CHLORIDE CONCENTRATION T ccu = 2 5 KR
CFe
1.25 2.5 5 10 12.5
46 71 117 169 190
= 30'
ccu = 2 5 Cw
KR
15 20 30 35
207 241 289 317
..
...
ONLY N = 0.076 Chlorides Ccu = 12.5 CF~
2.5 12.5 20 25 37.5
KR
61 174 224 241 300
CC" = 37.5 CF~ KR
2.5 12.5 37.5
.. ..
55 148 286
.. ...
The curve for cupric chloride alqne may scarcely be distinguished from the horizontal axis; therefore, on the assumption that the joint effect of iron and copper salts is purely additive, the curve for this joint effect should coincide with the curve for ferric chloride alone. It will be noted, however, that the presence of only a small amount of copper salt causes a marked acceleration. A.large concentration of copper, as shown by Curves I1 and 111, has a smaller effect; it is therefore evident that some concentration of copper in the vicinity of 2.5 milligram-atoms per liter is more effective than any other. From these results we may exclude the hypothesis that the copper salt is activated by the iron.
Effect of a Varying Concentration of Copper with a Fixed Concentration of Iron That there is a maximum effective concentration of copper salt is better illustrated by the results shown in Figs. 2 and 3 (Tables I11 and IV). The curves represent the variation in the rate of reaction caused by varying concentration of copper salt in the presence of different fixed concentrations of iron salt. A relatively minute amount of copper has a marked effect, but in each case a maximum rate of reaction is obtained with a concentration of only 1 milligram-atom per liter. Beyond this point, the rate continually decreases with the addition of more copper, an effect which is quite analogous t o that caused by the addition of sodium chloride. The maximum effective concentration of copper appears t o be independent of the concentration of iron salt.
2516
VOl. 45
VAN L. BOHNSON AND A. C. ROBERTSON
TABLE 111 EFFECT OF A VARYING CONCENTRATION OF COPPER CHLORIDE ON THE RATEOF REACTION
T CFe = 2.5
ccu 0 0.05 0.2 0.375 0.5 1.o 2.5 12.5 37.5
...
I . .
CFe 3 ~.
KR
19.5 44 63 86 88 68 71 61 55
.. *.
..
...
N
= 30'
ccu 0 0.05 0.1 0.2 0.25 0.75 1.0 1.25 2.0 2.5 4.0 40
= 0.076 Ca. __
10 KR
ccu 0 0.1 0.2 0.3 0.5 0.875 1.5 2.5 5.0 30
100 135 146 159 165 174 174 173 173 169 166 133
-
e . .
20 KR
177 224 233 238 243 244 244 241 240 191 e . .
...
. . I
TABLE IV
EFFECT OF A VARYING CONCENTRATION OF COPPER SULFATE ON THE RATEOF R ~ A C T I O N Ai = 0.076 Cre = 7.6
ccu 0 0.1 0.3 0.4 0.6 1.0 2.0 4.0
...
KR
21 76 98 107 113 122 120 118
....
T = 30' N = 0.076 C F ~= 3 0 4 CCU KR
0 0.1 0.2 0.3 0.5 1.0 2.0 3.0 5.0
81 210 255 282 301 316 317 314 291
= 0.038 CF= ~ 7.6
A7
CCU
KR
0 0.1 0.3 0.5 1.0 1.5 2.0
35 126 166 184 197 202 202
... e . .
... I . .
That the effect is common to both chloride and sulfate systems is evident; the maximum concentration oi copper in each case is of the same order of magnitude, although ferric sulfate is a much less active catalyst than ferric chloride. It is also evident (Curve 11, Fig. 3, and last column, Table IV) that the maximum effective concentration of copper is not altered by a large decrease in the concentration of the free acid present. These results lead to the conclusion that the iron salt is activated by the presence of a copper salt. Conditions Affecting the Extent of Promotion Promotion Factors.-It has been shown that for any given concentration of iron, a maximum rate of reaction is obtained with approximately the same absolute concentration of copper in all cases. By dividing the maximum rate by the normal rate for the given concentration of iron, a ratio is obtained which may conveniently be termed the "promotion
NOV.,
1923
2517
I
PROMOTER ACTION IN HOMOGENEOUS CATALYSIS.
factor;” this factor represents the extent of promotion under given conditions. Effect of Concentration of Catalyst.-The promotion factors for different concentrations of iron salt can be calculated from the values obtained
0
SIjstem.
01
1 2 Mg. atoms of Cu per liter Fig. 2.-Effect on reaction rate of fixed iron with varying copper concentration 0
0
1 2 Mg. atoms of Cu per liter Fig. 3.-Effect on reaction rate of fixed iron with varying copper concentration
by interpolation from the curves of Fig. 1 and Fig. 4. The latter, for which the data are shown in Table V, represents the promotion process when sulfates of iron and copper are used (in the presence of free sulfuric acid). TABLE V EFFECT OF A VARYING CONCENTRATION OF FERRIC SULFATE Cwe. . . . . . . . . . . . . . 3.8 KR: no Cu . . . . . . . . . KR: Cc, = 1.0.... 72.5
VARIATION OF CF~
1 2 4 7 10
7.6 21
15.2 22.8 30.4 38.0 60.8 43 60 81 156 200 263 320 366 ... TABLE VI PROMOTION FACTORS WITH CONCENTRATION OR IRONSALT
...
..
Promotion factors Chlorides
5.0 3.6 2.7 2.1 1.7
Sulfates
8.0 7.0 6.7 6.0 6.6
CFe
15 20 25 30
40
Promotion factors Chlorides Sulfates
1.5 1.3 1.2 1.2 1.17
4.9 4.6 4.2 4.0 I.
2518
VAN
VOl. 45
L. BOHNSON A N D A. C. ROBGRTSON
I n this case, as is shown also in the last column of Table V I , the promotion factor is much larger than in the chloride system, although the normal rate is
7.5
'
$\,
\
;O\
~
-\
o\o
t
8
5.0
3 .-8 jO/ & 2.5
o
pp\3 .
o
1
\ \\
0
'
o-030-o
Fe C l
0
NOV., 1923
PROMOTER ACTION IN HOMOGENEOUS CATALYSIS.
2519
I
of catalyst is decreased, thus indicating that in very dilute solutions of the catalyst the mechanism of promotion is independent of the nature of the anion. It may be noted here that the promotion factor is the same when ferrous sulfate is the catalyst as when ferric sulfate is used, in dilute solution. This fact is in accord with the initial 30, rapid oxidation of ferrous sulfate, discussed in the previous paper. Effect of Concentration of Free Acid.-The effect of acid on the promoter action has already been partly shown in Curve 11, Fig. 3, where it is seen that in the sulfate system the maximum effective concentration 2 ~ , of copper is not changed by a decrease in the acid concentration; it can be shown that the promotion factor for Cp, = 7.6 is 5.8 whether the acid concentration is 0.038 N or 0.076 N. Similarly, for the chloride system, 101 -0the promotion factor also remains I Fe CI, alone I unaltered when the concentration of LI Fe CI,. 2.5 mq.-otom Cu free acid is changed from 0.076 N to I N-0.038 0.038 N. This is shown by the data of Tables VI1 and VIII, and in the I curves of Fig. 6. The promotion fac0: 10 Mg. atoms of Fe per liter tors obtained by interpolation from the latter are compared in Table VIII Pig. 6.-Promoter effect in the presence of decreased acid concentration with those obtained from Fig. 1. The fact that free acid has the same effect on the accelerated reaction that it has in the presence of iron salt alone, further confirms the hypothesis that the iron salt is the primary catalyst and the copper salt the promoter.
I
'
TABLE VI1
EFFECTOF DECREASED ACID CONCENTRATION ON THE FERRICCHLORIDE-COPPER CHLORIDE CATALYSIS
T
Cwe.. .............
2.5
CC"..
0
.............
K R. . . . . . . . . . . . . . . . 39.5
=
5 0 89
30 N = 0.038 12.5 17.5 0 0 216 293
1.25 2.5 96
7.5 2.5 275
10 2.5 317
TABLE VI11
EFFECT OF DECREASED ACID CONCENTRATION ON PROMOTION FACTORS cp,
2 4
Promotion factors N . i . 0.076 N = 0.038
3.6 2.7
3.5 2.7
CF~
7 10
Prontotion factors N = 0.076 !V = 0.038
2.1 1.7
2.1 1.8
2520
VOl. 45
L. BOHNSON AND A . C. ROBERTSON
VAN
Effect of Temperature.-The effect of temperature on the extent of promotion is also negligible. It is shown in Table I X that the promotion factor in the chloride system remains constant, even though the reaction rate is greatly increased by raising the temperature. That the temperature coefficient is unchanged by the presence of copper is shown by Table X ; this is again evidence for the hypothesis that the copper salt present activates the iron salt. TABLEIX
EFFECT OF TEMPERATURE O N PROMOTION FACTORS R, = normal rate, with ferric chloride; R,
=
rate when copper is present; P = promo-
tion factor -T CF~
2.5 5 10
-T
= 40°-
Rs
Rp
75 171 326
275 417 587
-7-
= 30O-7
p
Rn
RP
p
Rn
3.7 2.4 1.8
19.5 46 98
71 118 173
3.6 2.6 1.8
4.1 10.6
...
= 200-
Rp
P
15.1 25.5
3.7 2.4
..
...
TABLE X EFFECT on PROMOTER ON TEMPERATURB COBPFICIENTS Temperature coeficientsCPO
40130
2.5 5 10
3.85 3.72 3.47
F e alone
30/2
40/30
4.75 4.34
3.87 3.54 3.40
*.
-
F e = CII
30/20
4.70 4.63
..
Effect of the Anion.-The effect on the reaction rate of the addition of various potassium salts to the reaction mixture when iron salts alone are
T =
TABLE XI EFIWCT OF NEUTRAL SALTS ON THE PROMOTER ACTION 30' N = 0.076 CFe(asFeCls) = 10.0 ccu = C,
=
Maximum rate, C,
0 1
5 10 15 16 18 20 25 30 50 75 100 150 400 1880
1.1
concn. of neutral salt in millimoles per liter KR
K CI 176
KNO
KBOd
KaPOa
176
176
159
176
... ... ...
145 132 114
176 150 76 33
...
... ...
175
105
... ... . I .
...
... 153 142 138 132 124 86
...
*..
... ... 171 ... 165 ... 157 136
84 62 44
...
32 25
...
..
1.2 0.8 0.6
...
... ... ... ... ... ...
...
Nov., 1923
PROMOTER ACTION I N HOMOGENEOUS CATALYSIS.
I
2521
the catalysts, has been shown in a previous paper. The effect of such salts when copper is present as a promoter is shown in Fig. 7 , the data being given in Table XI. A comparison of these results, plotted logarithmically, with those presented in the previoils paper shows that the rate due to molecular salt is accelerated] as well as the rate due to iron ions.
The Mechanism of Activation It was shown in the previous paper2 that the decomposition of hydrogen peroxide by iron salts alone is due to the formation and subsequent reduction by hydrogen peroxide of an intermediate compound, probably H2Pe04. The rate of evolution of oxygen from a given concentration of hydrogen peroxide is proportional to the concentration of the intermediate compound, 0 100 200 300 400 Millimoles per liter which is in turn dependent upon the concentration of iron salt and Fig. 7.-Bffect of neutral salts on the promoter action
of acid, and differs notably, depending upon whether the chloride] sulfate or nitrate of iron is used. In other words, the rate of reaction measured by the gasometric method can be no faster than the slower of the intermediate reactions] whether this be the formation of ferric acid, or its reduction by hydrogen peroxide, or some step in these processes. It is conceivable that the promoter, which is thus a secondary catalyst] might accelerate either one or both of the intermediate processes. Two equal portions of a solution of barium ferrate in acetic acid were allowed to react with equal quantities of hydrogen peroxide. To one of the samples a very little copper acetate was added; measurements of the oxygen evolved showed that this one reacted much faster than the one to which no copper salt was added. This is an indication that the copper accelerates the reduction of the intermediate compound. This experiment is not sufficient to explain the noticeable fact (compare Fig. 1 with Fig. 4) that ilthough the rate for a given iron concentration is much lower when the sulfate is used, the maximum “promoted’) rate is very close to that obtained when the chloride is used. If, however, we assume that the copper salt also accelerates the formation of the intermediate compound, the facts are more easily understood. It is probable] therefore, that the copper acts as a secondary catalyst] accelerating both
2522
ARTHUR W. THOMAS AND ALEXANDER FRIEDEN
Vol. 45
the intermediate reactions. The exact mechanism of this secondary catalysis is not yet clear.
summary 1. A small amount of copper salt greatly accelerates the decomposition
of hydrogen peroxide by iron salts in acid solution. Other metallic salts do not appear to have this “promoter effect.” 2. For a given concentration of iron salt, either sulfate or chloride, there is a maximum rate of reaction with about 1 millimole per liter of added copper salt. This maximum effective concentration of copper appears to be independent of the concentration of iron salt. 3. The extent of promotion (measured by the “promotion factor,” which is the ratio of the maximum rate to the normal rate for the iron alone) is independent of the acid concentration and temperature. 4. The extent of promotion is much greater for ferric sulfate than for ferric chloride, although the maximum rate obtained in each case is the same. The promotion factors for the two systems approach each other as the concentration of iron approaches zero. 5 . The facts are best explained by the following assumptions regarding the mechanism of promotion. (a) The iron salt is the primary catalyst, by virtue of its ability to form an intermediate compound. (b) The copper salt is a secondary catalyst, accelerating both the formation and subsequent decomposition of the intermediate product. The effect of the secondary catalyst is dependent upon its concentration only for very small quantities, a maximum effective concentration being noticeable. The mechanism of this secondary catalysis has not yet been completely explained. MADISON, WISCONSIN [CONTRIBUTION gROM THE CHEMICAL LABORATORIES, COLUMBIA UNIVERSITY, NO. 4201
FERRIC SALT AS THE “SOLUTION LINK” IN THE STABILITY OF FERRIC OXIDE HYDROSOL‘ BY ARTHURW. THOMAS AND ALEXANDERFRIEDEN RGCEXVGD JWNB 13,1923
Ferric oxide hydrosol was first prepared by Arnold M a w 2 It was regarded as a chemical compound3 until the introduction of the process of dialysis. Pdan de St. Gilles4 and Thomas Graham6 looked upon it as 1 This work was completed in February, 1922, and is adapted from the dissertation submitted by Alexander Frieden in partial fulfilment of the requirements for the degree of Doctor of Philosophy in the Faculty of Pure Science, Columbia University, 1922. 2 Maus, Ann. Phys. Chem., 11,75 (1827). 3 Soubcrain, Ann. chim. phys., 44, 325 (1830). Rose, Ann. Phys. Chem., 24, 301 (1832). SchBnbein, ibid., 39, 141 (1836). Scherer, ibid., 44, 453 (1838). Berzelius, “Lehrbuch der Chemie,” 1845, vol. 111, p. 555. Pean de St. Gilles, J . prakt. Chem., [l] 66, 137 (1855). 5 Graham, Phil. Trans., 161, 183 (1861).